Drug-drug interactions (DDIs) are an important public health problem, especially in the elderly, who are the largest consumers of medications. Nearly all existing knowledge about the clinical importance of potential DDIs comes from pharmacokinetic studies or case reports, both of which have limited utility for identifying clinically important DDIs. As a result, there is a great deal of disagreement among standard DDI reference texts regarding the clinical importance of most potential DDIs. Systematic efforts such as drug utilization review programs and computerized physician order entry that seek to improve medication use depend entirely upon a valid set of clinical rules, and are thus severely hampered by this knowledge gap. To address these critical gaps in existing knowledge of the clinical importance of DDIs, the investigators will perform controlled epidemiologic studies of high-priority potential DDIs that have been selected based on a formal process involving consultation with an Expert Panel of clinicians and clinical pharmacologists. We will use and expand upon an existing, very large pharmacoepidemiologic database consisting of Medicaid data from five large programs, supplemented by Medicare data to ensure complete capture of outcomes in the elderly, together with the General Practice Research Database from the UK. We will examine the validity of our outcome measures using medical record review. Our research strategy will include nested case-control and case-crossover designs, and will yield relative risks, risk differences, and number needed to harm (NNH). Patient and treatment factors associated with developing the adverse outcome of interest will be studied. The results of this study will have important implications for public health, clinical care, and the elucidation of the pharmacology of DDIs, and will serve as a basis for future studies examining factors not recorded in administrative and medical records data.